A1119
Title: Separable effects under semicompeting risks
Authors: Jih-Chang Yu - Academia Sinica (Taiwan) [presenting]
Yen-Tsung Huang - Academia Sinica (Taiwan)
Abstract: The separable effect has recently been proposed to study the causal effects under the setting of competing risks. The separable effect approach is extended to the semi-competing risks involving a primary outcome and an intermediate outcome. The exposure into two disjoint components is decomposed: the first component affects the primary outcome directly, i.e., direct effect and the other affects the primary outcome through the intermediate outcome, i.e., indirect effect. Under such effect separation, the identification formula of counterfactual risk derived for semi-competing risks is a function of cause-specific hazards and transition hazards of multistate models. It can be reduced to the formula for competing risks as a special case. Both nonparametric (NP) and semiparametric (SP) methods are proposed to estimate the causal effects and study their asymptotic properties. The model-free NP method is robust but less efficient for confounder adjustment; the model-based SP method flexibly accommodates confounders by treating them as covariates. Comprehensive simulations are conducted to study the performance of the proposed methods. Finally, the proposed methods are applied to characterize how hepatitis C infection affects the incidence of liver cancer through liver cirrhosis.
